51 minutes ago, abrams_tank said:

Those users are a pretty big segment of MSFS users now.

Would you mind giving us some hard data to back up your claim?

28 minutes ago, Lucky38i said:

IMO Live weather should be consistent with irl weather. If the METAR is wrong, the live weather shouldn't reflect that. That's just bad/late reporting and I imagine that occurs sometimes irl as-well which is why it helps to also read the TAF and crosscheck the METAR time with the current time. It's not iiregular for the METAR reported 30 minutes ago to be clear skies, then suddenly there's a downpour. MSFS should show the downpour, not a METAR report from 30 min ago.

How would MSFS get data for real life weather though, in real time, for every airport around the world?

Edited March 29, 20233 yr by abrams_tank

18 minutes ago, asanosho said:

Would you mind giving us some hard data to back up your claim?

Source: https://download.navigraph.com/docs/flightsim-community-survey-by-navigraph-2022-final.pdf

64.5% of flight simmers responding to the question "Have you flown online in the past 12 months?", said "Yes." [See p.68]

Of those that responded yes, approximately 85% have used the big 3 online networks (VATSIM ~60%, IVAO ~20%, Pilot Edge ~5%) [See p.70]

Therefore, 64.5% * 85% = 54.8% of survey respondents use a major online network when flight simming.

 

While it's difficult to ascertain just how many flight simmers responded to these two questions (the survey was only completed by 63% of the 25,427 people that started the questionnaire and the online-related questions appeared towards the end of the survey), with nearly 55% of respondents using large online networks when flying, it's a big enough number to make MS/Asobo take notice of the METAR weather request.

Just to broaden the discussion a bit, it is very much in Microsoft's strategic interest to develop the concept of a "world simulator" that goes well beyond flightsim for other applications. These might include providing almost real time environmental input for AI apps that have nothing to do with games. Flight sim is small potatoes in the larger context of such possibilities.

The discussion of global model vs. METAR-based weather seems to come up on a regular basis here on AVSIM. Local weather reports AND global model forecasts are critical to aviation in the real world, and flight sims that limit their weather simulation to one or the other are deficient in terms of providing even a quasi-realistic picture of the Earth's atmosphere. MSFS could certainly do better, but that's going to come at significant computational cost and complexity. In the meanwhile, let's enjoy what we have because it's an amazing step up from where we have been.

5 minutes ago, F737MAX said:

64.5% of flight simmers responding to the question "Have you flown online in the past 12 months?", said "Yes." [See p.68]

Yes, but the survey respondents were themselves very likely a highly unrepresentative population that was dominated, for example, by Navigraph users. Compared to the 10 million plus users who at one time or another have accessed MSFS, the survey respondents must be a relatively modest number, or the people behind Navigraph must be very rich indeed.

19 minutes ago, abrams_tank said:

How would MSFS get data for real life weather though, in real time, for every airport around the world?

MeteoBlue, MSFS's weather coverage supplier provides real-time weather with global coverage at varying altitudes, outlined in their API spec https://www.meteoblue.com/en/weather-api#info. So I'm confused by your question, as demonstrated by some of the initial posts in combat to OP's original post, MSFS is already doing this.

10 minutes ago, F737MAX said:

with nearly 55% of respondents using large online networks when flying, it's a big enough number to make MS/Asobo take notice of the METAR weather request.

Would you call less than 9000 people a pretty big segment of MSFS users?

15 minutes ago, Lucky38i said:

MeteoBlue, MSFS's weather coverage supplier provides real-time weather with global coverage at varying altitudes, outlined in their API spec https://www.meteoblue.com/en/weather-api#info. So I'm confused by your question, as demonstrated by some of the initial posts in combat to OP's original post, MSFS is already doing this.

Meteoblue weather isn't real time, it's predicted weather using an AI learning model.

Edited March 29, 20233 yr by Tuskin38

16 minutes ago, Lucky38i said:

MeteoBlue, MSFS's weather coverage supplier provides real-time weather with global coverage at varying altitudes, outlined in their API spec https://www.meteoblue.com/en/weather-api#info. So I'm confused by your question, as demonstrated by some of the initial posts in combat to OP's original post, MSFS is already doing this.

What @tuskin38 said. MeteoBlue's data isn't real time.

5 minutes ago, jrw4 said:

Yes, but the survey respondents were themselves very likely a highly unrepresentative population that was dominated, for example, by Navigraph users. Compared to the 10 million plus users who at one time or another have accessed MSFS, the survey respondents must be a relatively modest number, or the people behind Navigraph must be very rich indeed.

By its very nature, anyone who takes the time to answer *any* survey is likely to be an unrepresentative population, no matter who conducts it.

Be careful with the 10m number. As you say, it's people "who at one time or another have accessed MSFS". We don't know how they are counted, nor how many of those 10m use MSFS on a regular basis today.
Looking at Steam Charts, all-time peak users of MSFS was 61,775. Now peak numbers are averaging at less than 20% of that figure. Could 1.6m people still be using MSFS at peak time?  🤷‍♂️

 

3 minutes ago, asanosho said:

Would you call less than 9000 people a pretty big segment of MSFS users?

This site suggests VATSIM had 10,262 flights recorded in the 24-hour period on Saturday 18th March, which appears to be standard for any Saturday or Sunday in March.
This site suggests that there are 38,389 on the IVAO network. Though again, caution is needed. What time period does this relate? Does this count individual accounts or sessions? etc.

 

Whether representative of MSFS users or not, the Navigraph survey shows a not insignificant number who use online networks when flying.
The real take-away point of this is ~1.55m of (perhaps) 1.6m regular users of MSFS don't care either way about live weather and METARS, but for somewhere between 10,000 and 50,000 users, it matters and they created enough noise to move Asobo to create METAR-based live weather.

36 minutes ago, F737MAX said:

Whether representative of MSFS users or not, the Navigraph survey shows a not insignificant number who use online networks when flying.
The real take-away point of this is ~1.55m of (perhaps) 1.6m regular users of MSFS don't care either way about live weather and METARS, but for somewhere between 10,000 and 50,000 users, it matters and they created enough noise to move Asobo to create METAR-based live weather.

It is somewhat inaccurate to say that Live Weather is “METAR-based”. METAR observations do play a role, but they are not the entire foundation of the MSFS weather system.

Airport surface temperature, wind direction and speed come from METAR, (if available), but those values transition to the MeteoBlue model values at 1500 feet AGL. From there up to the stratosphere, those values come only from the MB model.

Clouds and visibility are another matter. I do not know where or how those transition between METAR and MB model. I am certain though, that all mid and high-level clouds come from the model, not from METAR. There was a recent post on the MSFS forums where a user was complaining that Live Weather was “wrong” because he had a solid overcast at EGLL, when the (automated) EGLL METAR showed CLR. But, other users who live near London posted r/w photos that showed that there was indeed a solid overcast at that time. I don’t know what the cloud bases were, but they apparently were coming from the model, not from  METAR.

I recently did a flight from Los Angeles to Honolulu while one of the recent infamous “atmospheric river” storms was approaching the west coast. The real-time satellite showed heavy cloud cover and rain extending about 800 miles west of California, before gradually transitioning to partly cloudy closer to Hawaii. That is exactly what Live Weather provided, and those clouds would have been exclusively from the MB model, as there are no METAR-reporting stations at all for almost 2000 miles along the route I flew.

I say this only because some MSFS users are (I think) convinced that Asobo scrapped the entire original Live Weather system and believe that all in-sim weather comes only from METAR - which is definitely not the case.

 

2 hours ago, abrams_tank said:

What @tuskin38 said. MeteoBlue's data isn't real time.

That's a fair point. I think it goes to show that the ML Modelling is relatively accurate to current real-world weather no? I'd prefer that than having weather based on old METARs

2 hours ago, F737MAX said:

Be careful with the 10m number. As you say, it's people "who at one time or another have accessed MSFS". We don't know how they are counted, nor how many of those 10m use MSFS on a regular basis today.

Yes indeed, one does need to be careful with that number which is why I was being careful in how I stated it.

BTW, why would one decide that only online pilots cared whether or not the depiction of airport weather in MSFS adhered closely to that reported in the METARs? To my recollection, no previous airport weather depiction in any simulator, whether provided by default or by a weather add-on, ever depicted weather based on a model forecast rather than METARs. If someone remembers any that did so, please let us know. Flight simmers participating in virtual airlines, various "pilot career" add-ons, etc., will likely want to fly in an atmosphere that reflects current (and sometimes historical) real world conditions. Ditto for those of us who are or were real world pilots or who like to use widely available free and paid sources of global weather. There's nothing wrong with preferring the depiction that can be derived from a model atmosphere that ensures smooth spatial and temporal conditions, but that's one of those trade-offs which we need to make in flying in desktop simulator.

 

 

8 minutes ago, jrw4 said:

BTW, why would one decide that only online pilots cared whether or not the depiction of airport weather in MSFS adhered closely to that reported in the METARs? To my recollection, no previous airport weather depiction in any simulator, whether provided by default or by a weather add-on, ever depicted weather based on a model forecast rather than METARs. If someone remembers any that did so, please let us know. Flight simmers participating in virtual airlines, various "pilot career" add-ons, etc., will likely want to fly in an atmosphere that reflects current (and sometimes historical) real world conditions. Ditto for those of us who are or were real world pilots or who like to use widely available free and paid sources of global weather. There's nothing wrong with preferring the depiction that can be derived from a model atmosphere that ensures smooth spatial and temporal conditions, but that's one of those trade-offs which we need to make in flying in desktop simulator.

The common refrain heard was that by using a weather model forecast, rather than old-style METAR-based depictions, the weather in MSFS could be wildly different from the weather conditions in use on online networks like VATSIM or forecast by Virtual Airlines' preflight tools.
Complaints of having to land with significant tailwinds and visibility differences when compared to the what the online network and METARs were reporting.

If the sim's weather was replicating similar conditions to real reported weather, e.g. if the model forecast had more frequent updates throughout the day to reflect changes from the original forecast, then I don't think there would've been such a clamour for change.
So long as the winds in MSFS allowed approaches from the same direction as the online controllers were using and depicted vaguely appropriate conditions when FG / RA / TS / SN was present, I don't think people would have cared as much—though I admit that this point is a wild guess on my part.

I suppose it's a bit like the lightning in the sim. There were plenty of justifiable complaints from users when lightning appeared in CAVOK or clear conditions, yet now it's rarely seen.

The OP misunderstood what was happening and got corrected soooo why are we still talking about this? lol